International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

A Review of the Variations of Optical Remote

Sensing Conditions over Estonia in 1958-2011
Kalju Eerme*1, Margit Aun1, 2
1

Tartu Observatory, Travere 61602, Tartumaa, Estonia

2
University of Tartu, Tartu, Estonia
kalju@aai.ee
concerned with choosing the sites for astrophysical
observatories in Caucasus and Crimea are described in
recently issued paper [1] and these for other sites in several
earlier publications [2-4]. Estimation of the availability of
relevant conditions for observing the Earth from space may
be considered as the remote sensing climate. One of
important astroclimatic quantities is the number of nights
suitable for observations. In optical remote sensing the
number of days suitable for remote sensing may be
considered as a similar indicator. The remote sensing
conditions are far from being equal at different geographical
regions. However, the remote sensing results are necessary
also at regions where the weather conditions are rather bad.
Some problems of remote sensing need relatively high
frequency of coverage or collecting data in certain
phenological phases of plant cover. Some objects are better
distinguished in the presence of snowcover.

Abstract- The major restricting factor for space-born optical

remote sensing at moderate and high latitudes is cloudiness.
Significant intraseasonal and interannual differences in cloud
amount and dominant cloud types are met. Cloud-free episodes
are infrequent. The study of cloud restriction over Estonia was
performed using hourly cloud detection data from TartuTravere meteorological station (5816N, 2628E, 70 m a.s.l.)
in 1958-2011. Major features of sky coverage by almost
nontransparent low and medium level clouds are considered.
Monthly mean amounts of these clouds in October to February
are above 7 tenths. Maximum mean coverage, close to 8.5
tenths, is reached in November. In spring and summer the
mean coverage is usually between 5 and 6 tenths. The spaces of
sky free from two lower level clouds often are partly covered by
semitransparent cirrus clouds. Monthly relative coverages by
cirrus clouds were studied statistically for March to September.
Relative monthly mean cirrus cloud coverage of free spaces
above one tenth was the largest, 0.40, in April and the smallest,
about 0.30, in July. Overall mean coverage for March to
September was 0.35. The days suitable for optical remote
sensing were defined as those with cloudless conditions within
2 hours from noon when remote sensing activities usually are
performed. The monthly averages and extremes of those days
are presented as well as their time evolution in four seasons
during 1958-2011. The probability of cloudless conditions is the
highest in March when nearly 4 suitable days are met on
average. The probability decreases from March to November
and since then the conditions start to improve. In November the
remote sensing conditions are available in less than half of the
studied years and in March in about 90 percent out of all
considered years. Monthly maxima of suitable days in all
months were above 5.

When the sky is overcasted by low and medium level

clouds, the optical remote sensing of land surface is
impossible or restricted to the free spaces between clouds.
Within spaces free from two lower level clouds the data
quality may be restricted by partly transparent upper level
clouds. Finally, certain restrictions of the reliability of results
may occur even in fully cloud-free conditions. The spatially
and temporary varying aerosol optical depth (AOD) and its
distribution by wavelength may be a reason of significant
uncertainties in atmospheric correction of data.
The remote sensing conditions are closely related to the
availability of solar irradiance on what we have published
several papers, both on the availability of broadband
irradiance as well as on the availability of ultraviolet
irradiance [5-12]. In the present paper the results of statistical
investigation of interannual and intraseasonal variations of
the amounts of almost nontransparent low and medium level
clouds as well as of the coverage by cirrus clouds of sky
spaces free from two lower level clouds are presented and
discussed. The considered period 1958-2011 includes 54
years. The results are based on hourly cloud visual detection
data collected at Tartu-Travere meteorological station
(5816N, 2628E, 70 m a.s.l.) located not far from the
weighted geographical centre of Estonia. Total area of
Estonia is about 45 000 km2. The landscape around consists
of arable land, grassland and forest. Cloud detection has been
performed also at several other meteorological stations but
with lower time resolution. Comparison of data from
different stations has revealed that over some other parts of
Estonia the sky may be to some extent cloudier. Over islands

Keywords- Remote Sensing; Cloud; Solar Radiation;

Meteorological Conditions; Suitable Days for Optical Remote
Sensing

I.

INTRODUCTION

Optical remote sensing of the Earth surface from space

needs favorable atmospheric conditions for making the
activities possible and avoiding uncertainties related to
changing atmospheric transparency over scenes. A similar
problem was actual in astronomy before the earth remote
sensing era. Fast development in astrophysics since 1950s
and worsening of observing conditions in observatories
placed close to towns caused building of new astronomical
observatories at remote sites preferably offering more
cloudless and stable atmospheric conditions. Special
astroclimatic studies at candidate sites were carried out to
select the ones presenting most favorable atmospheric
conditions. The principal stages of astroclimatic studies

12

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

located in western part of territory the sky tends to be to some

extent less cloudy than at the Tartu-Travere meteorological
station [13]. The differences are not large and the conclusions
made on the basis of Tartu-Travere data can be considered
representative enough for the whole country.
II.

B. Main features of Sky Coverage by Low and Medium

Level Clouds
In previous treatments of the Tartu-Travere
meteorological station cloud data the separation was made
between total clouds and low clouds. Also the long-term
trends of dimming and brightening were analyzed [16, 17].
The contribution of middle level clouds together with that of
upper level clouds was included in total clouds [18, 19]. The
treatment and analysis were performed on the monthly level.
In the present work the monthly averaging of the daily
average values of sky coverage by the sum of low and
medium level clouds were supplemented by the averaging
additionally over ten day intervals in each month. In February
the last 10-day interval includes 8 or 9 days and in months
consisting of 31 days 11 days, respectively. A matrix of sky
coverage by sums of low and middle level clouds consisting
of 1944 points and exposing 36 ten-day intervals in each of
54 years is presented in Figure 1. One can see that the
summarized cloud cover at two lower levels is very variable.
The probability density distributions of ten-day averages over
the considered 54 years were in all cases negatively skewed.
The skewness varies in range from -0.003 to -2.05, being
most frequently around -0.60. Due to small negative
skewness almost in all cases the median of distribution is by
0.1 to 0.4 tenths larger than the mean. Both the sharper and
the flatter distributions than the Gaussian were found.
Maximum coverage of sky exceeding 8 tenths on average is
reached in November. Since then it decreases to the value
around 5 tenths in last decade of April. Approximately the
same average coverage persists until the middle of June when
it increases by 0.5 tenths and remains on the same level until
the middle of August. The year-to-year variations of ten-day
averages are large. In all months are met the episodes when
the total amount of low and medium level clouds is small.
Those situations present potentially higher probability for
successful remote sensing activities. The lowest ten-day
average amounts of low and medium levels clouds in March
to August were below or slightly above 1 tenth. Maximal
amounts during the same period were above 9 tenths in
March, April and August and around 8.5 tenths in June and
July. The lowest maximal ten-day average coverage by low
and medium clouds, 7.5 tenths, was found in May. In
September to February the minimal ten-day average amounts
of low and medium level clouds varied in range between 1.8
and 2.6 tenths. Maximal values at the same time reached 10
tenths. In the cold and dark part of year, October to February,
overcast conditions are met more frequently than during the
period March to September.

MATERIALS AND METHODS

A. Cloud Data
In the early 1950s the Tartu-Travere meteorological
station was specializing on solar radiation measurements. The
first attempts of recording sunshine duration were made at
Tartu since 1906 [14]. First regular measurements of solar
irradiance were performed in late 1930s and continued after
the World War II since 1950 [15]. Before 1965 the station
was based closer to Tartu than its present site at 20 km from
the town center. The landscape around was similar to that at
the present site. For solar irradiance measurements the
Yanishevski AT-50 actinometers and Savinov-Yanishevski
M-115 pyranometers were used until 1996 but were since
replaced by the Eppley Labor. Inc. pyrheliometers and Kipp
& Zonen pyranometers. The absolute accuracy of the
ventilated Kipp & Zonen pyranometers is about 2% and that
of the pyrheliometers 1%. In the case of older instruments
these uncertainties usually were doubled. In the past
intercalibration of sensors was regularly performed in
Voeikov Main Geophysical Observatory (St. Petersburg,
Russia), whereas now it is done in World Radiation Center
(Davos, Switzerland). Cloud data were collected as auxiliary
information for interpreting the solar radiation measurements.
Since July 1957 the hourly visual cloud detection at all
three basic altitude levels was performed being unique among
meteorological stations of the former Soviet Union. These
visual cloud observations containing no gaps since the
beginning are continued at present time when the station
belongs to the Baseline Surface Radiation Network (BSRN).
The clouds are detected half past each hour in local solar time.
The cloud amounts in tenths as well as the dominating cloud
types are noted down. From the beginning of cloud
observation the total coverage of sky as the whole as well as
the coverage at all basic levels was performed in tenths. As
the clouds on two lower levels are almost nontransparent we
consider their total amount separately from the upper level
clouds. Daily average sky coverage by low and medium level
clouds was calculated considering the data from observations
within the range 0.5 hours from the sunrise to 0.5 hours
from the sunset. Treating the daily cloud data we noted down
roughly also the changes in major cloud types during a day as
well as longer clear episodes. In warm and more sunny
summer half-year the clear episodes appeared with higher
frequency in early morning and late evening and much less
frequently around noon [12]. The reason is development of
convective clouds over the nonuniformly heated surface with
increasing solar elevation. In cold and dark winter half-year
the frontal overcast cloudiness dominates and clear episodes
do not manifest preferred time during a day.

C. Upper Level Clouds

When the sky is free from low and medium level clouds
there remains a possible restriction of remote sensing by the
cirrus clouds. Major features of cirrus cloud amounts on the
monthly level as well as year-to-year variations of their
amount at the study site in 1958-2003 were discussed in our
work [4]. Here the data from years 2004-2011 have been
added and a new survey was made. The results from the
extended data set are very close to the previous ones. The
representativeness of monthly cirrus cloud amount depends

13

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

on the number of days in the current month suitable for cirrus

detection and on the spaces free from low and medium level
clouds on these days. The days with the averaged value of
free spaces above 1 tenth were accounted in further cirrus
cloud detection. As was described in our previous work on
cirrus clouds, a stronger criterion, accounting for those days
with the average value of free spaces above two tenths, was
also applied. The numbers of days in month suitable for
cirrus detection occurred in most cases by two days smaller
for the stronger criterion. Also the monthly average free
spaces occurred by nearly 0.5 tenths larger. However, the
monthly cirrus coverages of free spaces did not show
differences exceeding 0.15 tenths. In the present work we
used only a threshold 1 tenth for the free spaces. Monthly
averages, StDev limits, maxima and minima for those days

are presented in Figure 2. The time evolution of their annual

sums during period March to September (214 days) in 19582011 is presented in Figure 3. The monthly averages of
spaces free from low and medium clouds with StDev limits,
maxima and minima are presented in Figure 4. The time
evolution of average free spaces for March to September in
tenths is presented in Figure 5. In this case the mean of
monthly average values over seven months was used for
characterization of each year out of the studied 54. Similarly
the relative coverages of free spaces by cirrus clouds were
treated and presented. Here is necessary to mention that here
we considered only the days with free spaces above one tenth
but in constructing the Figure 1 all days in month, because
the relative low and medium cloud amounts are
systematically larger.

Figure 1 Averaged over ten days amounts of low+medium level clouds in tenths during 1958-2011

The detected free spaces may be fully, partly or not

covered by upper level clouds. All these situations really
were met. We use the monthly relative coverage of free
spaces by cirrus clouds. This characteristic does not depend
on the size of free spaces. The episodes when the relative
coverage is small or even zero during several following days
were met as well as the episodes of prevailing large relative
coverages of free spaces. The monthly relative coverage of
free spaces by cirrus clouds depends on the proportions of
different relative coverages. The monthly mean cirrus cloud
coverages with the evolution of March to September
averaged relative coverage is presented in Figure 7.
Figure 3 Time evolution of March to September (214 days) annual total
number of days with daily average spaces in low+medium clouds above 1
tenth

Figure 2 Number of days in month with daily average free spaces in

low+medium clouds above 1 tenth. Monthly average, StDev limits, max and
min

Figure 4 Monthly average spaces free from low+medium clouds, in tenths.

Monthly average, StDev limits, max and min

14

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

spring season April to June, the summer season July to
September and the autumnal season October to December.
Time evolution of annual numbers of suitable days is
presented in Figure 8 and the time evolutions of seasonal
numbers in Figure 9.

Figure 5 Time evolution of March to September spaces free from

low+medium clouds, in tenths

D. Suitable Days for Optical Remote Sensing

Optical remote sensing depends on the sun as the sole
source of illumination. The sun-synchronous orbits are
preferred to cross the regions at the same solar time. Very
low and very high solar elevations are usually avoided. At the
study latitude the crossing time is within a few hours from the
noon. From these considerations the acceptable conditions for
remote sensing are when there are cloud-free conditions
around noon.

Figure 7 Time evolution of March to September relative coverage by cirrus

of spaces free from low+medium clouds

III.

RESULTS AND DISCUSSION

A. Low and Medium Cloud Intraseasonal and Interannual

Pattern

Using the local cloud observation data from our cloud

database the days presenting cloud-free conditions within 2
hours from the noon were selected as the suitable days for
remote sensing. The selection was done by visual checking of
the dataset where the cloud free episodes as well as the
overcast intervals and major cloud types in partly cloudy
conditions are presented. In most cases these days are almost
cloudless but the minor part is cloudy before or later of the
considered interval. The first two columns in Table 1 present
the monthly mean and maximal number of days suitable for
optical remote sensing. In the third column the number of
years out of 54 with no days suitable for the optical remote
sensing in current month is presented. The last column
presents the same in percent scale.

The pattern of averaged over ten days sums of low and

medium level cloud manifests significant interannual
variations.
TABLE I MONTHLY AVERAGES AND MAXIMA OF DAYS SUITABLE FOR
OPTICAL REMOTE SENSING IN 1958-2011. NUMBER OF YEARS OUT OF 54
AND PER CENT OF YEARS PRESENTING NO SUITABLE DAYS. THE SAME ON
SEASONAL SCALE
Month

Figure 6 Relative coverage of spaces free from low+medium clouds by

cirrus. Monthly average, StDev limits, max and min

The annual amounts of days suitable for optical remote

sensing were calculated as the sums of monthly numbers.
Similarly the numbers of those days in different seasons were
found. The winter season contains January to March, the

Average

Maximum

Years with No Suitable

Days
Number

Per cent

January

2.07

11

20.4

February

2.41

17

31.4

March

3.83

14

11.1

April

2.78

16.7

May

3.09

10

16.7

June

2.52

13

24.1

July

2.61

13

15

27.8

August

1.65

11

18

33.3

September

1.52

19

35.2

October

1.61

20

37

November

28

51.9

December

1.54

14

25.9

Winter

8.31

20

Spring

8.39

17

Summer

5.78

20

5.5

Autumn

4.15

13

3.7

The autumnal months, especially November and

December, are cloudier than the remaining ones. In October

15

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

been small. In May the range of variation in 1970-1995 was
notably larger than earlier and later. In June the interannual
variations have been the smallest among months in the year.

the range of variation of averaged cloud amount was between

3 and 9 tenths. Since 1996 the interannual variations tend to
be similar to those in 1958-1968 and small values appear
more frequently than in years 1969-1995. Since the third tenday interval of October, the range of interannual variations
decreases and remains between 6 and 8 tenths. In November
also the range of variation remains within two tenths Often
January and February are also heavily cloudy. In cold winters
with frequent domination of high pressure the monthly
average low and medium cloud amounts are significantly
smaller than in warm cyclonic winters. Small amounts
appeared in January more frequently in 1963-1987. In the
first two ten-day intervals of January the low and medium
cloud amounts have increased in last 25 years. In the third
ten-day interval no such tendency was found. In February the
smaller coverages appeared more frequently before 1985.
However, the overall lowest value in the third ten-day
interval was recorded in 2011.

In July the contrasting small and large average cloud

coverages cause the bimodality of distribution separating
cloudy and relatively fine weather years. In 1960s-1970s and
in recent 15 years the smaller averaged cloud amounts are
met more frequently than in period from late 1970s to the
middle of 1990s.
In sunny summers like those in 1960s to early 1970s and
since 1994 the autumnal cyclonic period causing increase in
cloud amount tends to begin later than in years 1976-1993
manifesting wetter and cloudier summers [6]. In the 13
sunniest summers the cloud amount has noticeably increased
after the first ten-day interval of August only in two cases. In
other 11 (85 %) cases the increase in daily average cloud
amount has happened after the first (6 cases) or second (5
cases) ten-day interval of September. Vice versa, in summers
exhibiting average or above average cloud amounts the sunny
periods tend to appear before the middle of July. Only in 7
cases of total 39 (18 %) the first or second ten-day interval of
September has been relatively sunny.
The quasiperiodic weather variations in different scales
are met in all seasons. They depend on the activity of
westerly flow and on development and persistence of
anticyclonic situations. The major contributors to the weather
variations in winter are the Arctic Oscillation (AO), the North
Atlantic Oscillation (NAO), the Arctic Polar Vortex or
circumpolar vortex strength, the Northern Hemisphere Winter
Storm Tracks intensity and the Atlantic meridional
overturning circulation (MOC). In the positive phase of AO
more of atmospheric mass is located over midlatitudes and
less over Arctic. Sea level pressure and the weather are
strongly correlated with the AO index only in AO Core
Region of strong westerly flow [20]. In the negative phase of
AO the pressure difference and the westerly flow are weaker.
A prolonged positive period in the smoothed AO index has
been observed in 1988-1995.

Figure 8 Annual numbers of days suitable for optical remote sensing

In March the coverage of sky by low and medium level

clouds decreases and the range of interannual variations
increases. In April the range continues to decrease. Relatively
small coverages were met more frequently before 1970 and in
two recent decades. During the two last ten-day intervals the
interannual variations in last 20 years are similar to those
before 1970. In May and June the range of year-to-year
variations is large. In very recent years, the variations have

Figure 9 Seasonal numbers of days suitable for optical remote sensing

16

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

The NAO as a dominant mode of the atmospheric

behaviour in the North Atlantic sector is a large-scale
alternation of atmospheric mass with the centers of action
near the Icelandic low and the Azores high. After 1980 an
eastward shift of action centers has been noticed. Often a
steady relationship between the NAO and key climate
parameters of the North Atlantic and Northern Europe is
expected. These relationships are really variable in time and
tend to be more complex than previously thought [21]. The
NAO in winter is confined to the Atlantic sector during low
solar activity, whereas it tends to have a wider hemispheric
AO like structure during high solar activity [22, 23].
Significant correlations between the heat transport in the
North Atlantic Ocean and atmospheric processes over the
Nordic Seas and the Eurasian continent have been found for
periods 25 years and longer [24]. The heat anomalies
propagate along the cyclone tracks towards northeast. The
oceanic influence on the North Atlantic climate has been
nonstationary [25]. The Northern Hemisphere surface
temperatures have shifted to the larger warming since 1970
[26] and since the late 1980s the cloud amounts in winter
have increased in regions that are exposed to the westerlies
[27]. The eastern Europe experienced an advanced annual
cycle of spring near-surface temperature in 1978-2010 [28].

C. Amounts of Suitable Days for Remote Sensing

The amount of days in month suitable for optical remote
sensing depends on weather conditions. The annual average
number of those days at the study site was 26.7 and their
maximal number was 50. Two most minimal annual values,
12 and 13, were recorded in recent decade, in years 2004,
2008 and 2009. The monthly average number of such suitable
days was the smallest, only one, in November and then
increasesd reaching maximum 3.8 in March. In April it
occurs somewhat smaller than in May and then decreases
toward the autumnal months. Maximum number, 14, was
also recorded in March. The second and third monthly largest
values, 13 and 11, were recorded in July and August,
respectively. In all months certain part of years present no
days suitable for remote sensing. In the Table 1 their number
out of 54 as well as the relative amount in percent scale are
presented. The smallest percent was found in March, which is
the most favorable month, by the remote sensing conditions.
The next two best months were April and May. In January
the amount of nonsuitable years is smaller than in December
and February. In winter months the numbers of days suitable
for remote sensing increases in the presence of cold high
pressure conditions. In warm cyclonic winters the favorable
conditions appear in very rare cases. In spring months April
to June at least one suitable day for remote sensing in year
was found. The maximum number of those days reaches 17
what is smaller of that in winter and summer but interannual
variation is the smallest among seasons. In winter high values
appear due to the large contribution of March. In the period
March to September the daily average spaces free from low
and medium clouds decrease. The major contribution to this
decrease comes from the hours around noon. As a result in
summer months the conditions for remote sensing are more
restricted than these in late winter-early spring and partly
even in early autumn. In October and in some few cases in
early November the conditions for optical remote sensing
have occurred better than in less cloudy on average summer
months despite that in more than 50% of years there have
been no days suitable for remote sensing in November. In
October the number of suitable days in 5 cases have reached
6 or 7 and in one exceptional case even 9. In July, August and
September the numbers above that level also have been very
exceptional.

B. Cirrus Cloud Distribution Within March to September

Period and Interannual
In days when the amount of low and medium level clouds
is small significant restrictions of optical remote sensing
conditions occur due to cirrus clouds. In some cases cirrus
clouds are met also in very small amounts. But in other
similar cases most of the free spaces in lower level clouds are
filled by cirrus. The monthly average coverage of sky spaces
free from low and medium clouds by cirrus was impossible to
detect in October to February when overcast days are met too
frequently. In the favorable months the number of such days
increased from relatively large value in March to the
maximum in April and then decreased to minimum in July
with following slight increase in August and September. The
interannual variations of relative coverage of free spaces of
sky by cirrus were the largest in March and then decreased to
minimum in July and August with the following increase in
September. The smallest monthly coverages by cirrus
remained in range 0.15 to 0.2, while the monthly largest
values decreased from 0.65 in March and April to 0.45 in
July with the following increase to 0.55 in September.
Standard deviations are to somewhat larger in March and
April and almost equal in other months.

IV. CONCLUSIONS
Over Estonia located at subpolar latitude in the zone
influenced by polar front and North Atlantic storm track the
conditions for optical remote sensing are moderate. The
monthly numbers of days suitable for remote sensing
activities tend to be relatively small and manifest significant
interannual variations. For each month a certain percent of
years with no such days was found in the period 1958-2011.
The percent was the highest (more than 50% of all studied
years) for November and the smallest (slightly exceeding 10)
for March.

Averaged over the period from March to September

values of coverage by cirrus clouds varies within range 0.26
to 0.45 with the mean value 0.35. Usual range of deviations
in separate years is within 0.05. Longer compact period of
positive deviation in years 1992-1996 is very probably
related to the increased concentration of cloud condensation
nuclei formed by falling sulphate aerosol after the Pinatubo
1991 eruption. The influence of the volcanic aerosol on cirrus
cloud microphysics is discussed in [29] and the transport
from tropics to midlatitudes in [30].

The most stable conditions for optical remote sensing

with the largest average number of suitable days in month
and the smallest percent of years presenting no suitable days
were met in March. The monthly average number of suitable

17

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

[8] K. Eerme, Variations of Total Solar Radiation and Estimated
Erythemal UV Doses in Estonia During 1953-2004, Annalen
der Meteorologie 41: 663-666, 2005.
[9] K. Eerme, U. Veismann, and S. Ltt, Proxy-based
reconstruction of erythemal UV doses over Estonia for 19552004, Ann. Geophys., 24, 1767-1782, 2006.
[10] K. Eerme, A. Kallis, U. Veismann, I. Ansko, Interannual
variations of available solar radiation on seasonal level in
1955-2006 at Tartu Travere Meteorological station, Theor.
Appl. Climatol., 101, 371-379, 2010.
[11] K. Eerme, Interannual and intraseasonal variations of the
available solar radiation, In: Solar radiation (Ed. by E. B.
Babatunde), InTech , Croatia, ISBN 978-953-51-0384-4, 3352, 2012.
[12] K. Eerme, I. Ansko, U. Veismann, Relationship between
daily relative sunshine duration and relative sum of direct
irradiance at Tartu-Travere meteorological station in 1967
2008, Estonian Journal of Engineering, 1, 1225, 2012.
[13] V. Russak, A. Kallis (compilers), H. Tooming (ed), Handbook
of Estonian solar radiation climate. EMHI, Tallinn, 384 pp (in
Estonian), 2003.
[14] A. Kallis, V. Russak, H. Ohvril, 100 Years of Solar Radiation
Report of the Eighth Session of the Baseline Surface Radiation
Network (BSRN), Workshop and Scientific Review (Exeter,
UK, 26.30 July 2004), WCRP Informal Report No. 4/2005,
C1.C4, 2005.
[15] H. Ohvril, H. Teral, L. Neiman, M. Kannel, M. Uustare, M.
Tee, V. Russak, O. Okulov, A. Jeveer, A. Kallis, H. Ohvril, E.
I. Terez, G. A. Terez, G. K. Guschin, G. M. Abakumova, E. V.
Gorbarenko, A. V. Tsvetkov, N. Laulainen, Global Dimming
and Brightening Versus Atmospheric Column Transparency,
Europe, 1906-2007. J. Geophys. Res. 114, D00D12,
doi:10.1029/2008JD010644, 2009.
[16] M. Wild, H. Gilgen, A. Roesch, A. Ohmura, C. N. Long, E. G.
Dutton, B. Forgan, A. Kallis, V. Russak, A. Tsvetkov, From
Dimming to Brightening: Decadal Changes in Solar Radiation
at Earths Surface, Science 308, 847-850, 2005.
[17] M. Wild, Global Dimming and Brightening: A review,
Journal
of
Geophysical
Research,
114,
doi:10.1029/2008JD011470, 2009.
[18] V. Russak, Trends in solar radiation, cloudiness and
atmospheric transparency during recent decades in Estonia,
Tellus, 42B, 206-210, 1990.
[19] S. Keevallik, V. Russak, Changes in the amount of low
clouds in Estonia (1955-1995), International Journal of
Climatology, 21, 389-397, 2001.
[20] J. Zhao, Y. Cao, J. Shi, Core region of Arctic Oscillation and
the main atmospheric events impact on the Arctic, Geophys.
Res. Lett. 33, L22708, doi: 10.1029/2006GL027590, 2006.
[21] E. I. Polyakova, A. G. Journel, I. V. Polyakov, U. S. Bhatt,
Changing relationship between the North Atlantic Oscillation
and key North Atlantic climate parameters, Geophys. Res.
Lett. 33, L03711, doi: 10.1029/2005GL024573, 2006.
[22] K. Kodera, Solar influence on the spatial structure of the
NAO during the winter 1900-1999, Geophys. Res. Lett. 30,
1175, doi: 10.1029/2002GL016584, 2003.
[23] K. Kodera, Y. Kuroda, A possible mechanism of solar
modulation of the spatial structure of the North Atlantic
Oscillation, J. Geophys. Res. 110: D02111, doi:
10.1029/2004JD0055258, 2005.
[24] T. Martin, E. Ruprecht, Decadal variation of the North
Atlantic meridional heat transport and its relation to

days decreased from nearly four in March to one in

November.
In separate years the favorable conditions for optical
remote sensing may exist in all months but with different
probability.
Despite the fact that the daily average amounts of low and
medium level clouds are smaller in spring and summer
months as compared to the autumnal and winter ones the
conditions for optical remote sensing in summer are not the
most favorable. In June and July about 2.5 suitable days in
month have been met and in August to October only about
1.5 such days in month on average. The amount of years with
no such days in month for the last interval exceeded 1/3 out
of all considered years. In June and July their relative
contribution was slightly above 1/4 out of all years.
In summer months the cloud amount increases around
noon due to the development of convection. The amounts of
convective clouds are smaller during the extended dry
periods and larger in wetter periods. The coverage of free
spaces of sky also tends to be smaller in extended dry periods
and larger in wetter ones.
Since the late 1980s the conditions for optical remote
sensing at the study site have rather worsened in winter and
spring as well as on annual level. Bad conditions have
frequently met also in summer and fall.
ACKNOWLEDGEMENTS

The work has been supported by the project 3.2.0801.110041 Estonian radiation climate of the Estonian Research
Council. The authors thank the Estonian Meteorological and
Hydrological Institute for the cloud data collected at the
Tartu-Travere meteorological station.
REFERENCES

[1] V. E. Panhuk and V. L. Afanasev, Astroclimate of Northern

Caucasus - Myths and Reality, Astrofizicheskii bulleten, 66,
2, 253-274, 2011.
[2] Sh. P. Darchia, Certain Problems in the Method of Studying
Astroclimate, Astronomicheskii Zurnal, 41, 147-155, 1964.
[3] L. R. Lisina, On the astroclimate of two points in the
Transcarpathian and Odessa regions of the Ukrainian SSR,
Astrometry and Astrophysics, 5, 94-101,1970.
[4] S. A. Ehgamberdijen, A. K. Baijumanov, S. P. Ilyasov, M.
Sarazin, Y.A. Tillayev, A. A. Tokoviniv, A. Zsiad, The
Astroclimate of Maidanak Observatory in Usbekistan, Astron.
Astrophys. Suppl. Ser, 145, 293-304, 2000.
[5] M. Aun, K. Eerme, I. Ansko, U. Veismann, and S. Ltt,
Modification of Spectral Ultraviolet Doses by Different
Types of Overcast Cloudiness and Atmospheric Aerosol,
Photochemistry and Photobiology, 87, 461-469, 2011.
[6] K. Eerme, U. Veismann, and R. Koppel, Variations of
erythemal ultraviolet irradiance and dose at Tartu/Travere
Estonia, Climate Research, 22, 245-253, 2002.
[7] K. Eerme, Changes in spring-summer cirrus cloud amount
over Estonia, 1958-2003, Int. J. Climatol., 24, 1543-1549,
2004.

18

International Journal of Remote Sensing Applications

Sept. 2012, Vol. 2 Iss. 3, PP. 12-19

atmospheric processes, Geophys. Res. Lett., 34, L04703, doi:

10.1029/2006GL028438, 2007.
[25] R. T. Sutton, D. L. R. Hodson, Influence of the Ocean on
North Atlantic Climate Variability 1871-1999, Journal of
Climate 16: 3296-3313, 2003.
[26] M. A. Ivanov, S. N. Evtimov, 1963: the break point of the
Northern hemisphere temperature trend during the twentieth
century, Int. J. of Climatol., 30, 1738-1746, 2010.
[27] A. Lehmann, K Getzlaff and J. Harla, Detailed assessment
of climate variability in the Baltic Sea area for the period 1958
to 2009, Climate Research, 46, 185-196, 2011.

[28] A. Schmidt, B. Law, C. Hanson and O. Klemm, Distinct

Global Pattern of Strong Positive and Negative Shifts of
Seasons over the Last 6 Decades, Atmospheric and Climate
Sciences, 3, 76-88, 2012.
[29] E. J. Jensen and O. B. Toon, The potential effect of volcanic
aerosol on cirrus cloud microphysics, Geophys. Res. Lett., 19,
1759-1762, 1992.
[30] O. A. Langford, T. I. O'Leary, M. H. Proffitt and M. H.
Hitchman, Transport of the Pinatubo aerosol to a northern
midlatitude site, J. Geophys. Res., 100, 9007-9016, 1995.

19